delete slice api (#48399)

python/paddle/fluid/contrib/layers/nn.py

@@ -36,7 +36,6 @@ from paddle.fluid import core
 from paddle.fluid.param_attr import ParamAttr
 
 from paddle.fluid.framework import Variable, convert_np_dtype_to_dtype_
-from paddle.fluid.layers import slice
 import paddle
 import warnings
 from paddle import _C_ops, _legacy_C_ops
@@ -1540,13 +1539,13 @@ def tdm_sampler(
 
         for layer_sample_num in neg_samples_num_list:
             end_offset = start_offset + layer_sample_num + positive_flag
-            layer_samples = slice(
+            layer_samples = paddle.slice(
                 out, axes=[1], starts=[start_offset], ends=[end_offset]
             )
-            layer_labels = slice(
+            layer_labels = paddle.slice(
                 labels, axes=[1], starts=[start_offset], ends=[end_offset]
             )
-            layer_mask = slice(
+            layer_mask = paddle.slice(
                 mask, axes=[1], starts=[start_offset], ends=[end_offset]
             )
 

python/paddle/fluid/dygraph/dygraph_to_static/convert_operators.py

@@ -29,7 +29,6 @@ from paddle.fluid.layers import (
 from paddle.fluid.layers import (
     assign,
     fill_constant,
-    slice,
     reduce_all,
     reduce_any,
 )
@@ -819,7 +818,7 @@ def _slice_tensor_array(array, start, end):
         return null_array
 
     def false_fn(array, start, end):
-        new_array = slice(array, starts=[start], ends=[end], axes=[0])
+        new_array = paddle.slice(array, starts=[start], ends=[end], axes=[0])
         return new_array
 
     new_array = cond(start == end, true_fn, lambda: false_fn(array, start, end))

python/paddle/fluid/layers/detection.py

@@ -1768,7 +1768,7 @@ def ssd_loss(
     target_label.stop_gradient = True
     conf_loss = softmax_with_cross_entropy(confidence, target_label)
     # 3. Mining hard examples
-    actual_shape = nn.slice(conf_shape, axes=[0], starts=[0], ends=[2])
+    actual_shape = paddle.slice(conf_shape, axes=[0], starts=[0], ends=[2])
     actual_shape.stop_gradient = True
     # shape=(-1, 0) is set for compile-time, the correct shape is set by
     # actual_shape in runtime.

python/paddle/fluid/layers/nn.py

@@ -110,7 +110,6 @@ __all__ = [
     'gaussian_random',
     'sampling_id',
     'sum',
-    'slice',
     'shape',
     'clip',
     'clip_by_norm',
@@ -6007,261 +6006,6 @@ def sum(x):
     return paddle.add_n(x)
 
 
-@templatedoc()
-def slice(input, axes, starts, ends):
-    """
-    This operator produces a slice of ``input`` along multiple axes. Similar to numpy:
-    https://docs.scipy.org/doc/numpy/reference/arrays.indexing.html
-    Slice uses ``axes``, ``starts`` and ``ends`` attributes to specify the start and
-    end dimension for each axis in the list of axes and Slice uses this information
-    to slice the input data tensor. If a negative value is passed to
-    ``starts`` or ``ends`` such as :math:`-i`,  it represents the reverse position of the
-    axis :math:`i-1` (here 0 is the initial position).
-    If the value passed to ``starts`` or ``ends`` is greater than n
-    (the number of elements in this dimension), it represents n.
-    For slicing to the end of a dimension with unknown size, it is recommended
-    to pass in INT_MAX. The size of ``axes`` must be equal to ``starts`` and ``ends``.
-    Following examples will explain how slice works:
-
-    .. code-block:: text
-
-        Case1:
-            Given:
-                data = [ [1, 2, 3, 4], [5, 6, 7, 8], ]
-                axes = [0, 1]
-                starts = [1, 0]
-                ends = [2, 3]
-            Then:
-                result = [ [5, 6, 7], ]
-
-        Case2:
-            Given:
-                data = [ [1, 2, 3, 4], [5, 6, 7, 8], ]
-                axes = [0, 1]
-                starts = [0, 1]
-                ends = [-1, 1000]       # -1 denotes the reverse 0th position of dimension 0.
-            Then:
-                result = [ [2, 3, 4], ] # result = data[0:1, 1:4]
-
-    Args:
-        input (Tensor): A ``Tensor`` . The data type is ``float16``, ``float32``, ``float64``, ``int32`` or ``int64``.
-        axes (list|tuple): The data type is ``int32`` . Axes that `starts` and `ends` apply to .
-        starts (list|tuple|Tensor): The data type is ``int32`` . If ``starts`` is a list or tuple, the elements of
-                it should be integers or Tensors with shape [1]. If ``starts`` is an Tensor, it should be an 1-D Tensor.
-                It represents starting indices of corresponding axis in ``axes``.
-        ends (list|tuple|Tensor): The data type is ``int32`` . If ``ends`` is a list or tuple, the elements of
-                it should be integers or Tensors with shape [1]. If ``ends`` is an Tensor, it should be an 1-D Tensor .
-                It represents ending indices of corresponding axis in ``axes``.
-
-    Returns:
-        Tensor:  A ``Tensor``. The data type is same as ``input``.
-
-    Raises:
-        TypeError: The type of ``starts`` must be list, tuple or Tensor.
-        TypeError: The type of ``ends`` must be list, tuple or Tensor.
-
-    Examples:
-        .. code-block:: python
-
-            import paddle
-
-            input = paddle.rand(shape=[4, 5, 6], dtype='float32')
-            # example 1:
-            # attr starts is a list which doesn't contain tensor.
-            axes = [0, 1, 2]
-            starts = [-3, 0, 2]
-            ends = [3, 2, 4]
-            sliced_1 = paddle.slice(input, axes=axes, starts=starts, ends=ends)
-            # sliced_1 is input[0:3, 0:2, 2:4].
-
-            # example 2:
-            # attr starts is a list which contain tensor.
-            minus_3 = paddle.full([1], -3, "int32")
-            sliced_2 = paddle.slice(input, axes=axes, starts=[minus_3, 0, 2], ends=ends)
-            # sliced_2 is input[0:3, 0:2, 2:4].
-    """
-    if in_dygraph_mode():
-        attrs = ()
-        starts_tensor = None
-        ends_tensor = None
-
-        if isinstance(axes, (list, tuple)):
-            axes = list(axes)
-            if len(axes) == 0:
-                raise ValueError(
-                    "Input axes should not be an empty list/tuple."
-                )
-            for i in range(len(axes)):
-                if axes[i] < 0:
-                    axes[i] = max(0, axes[i] + len(input.shape))
-                else:
-                    axes[i] = min(len(input.shape) - 1, axes[i])
-
-        else:
-            raise ValueError(
-                "Input axes must be a python list or tuple, but reveived {}".format(
-                    type(axes)
-                )
-            )
-
-        infer_flags = list(1 for i in range(len(axes)))
-
-        tmp_tensor_type = core.eager.Tensor
-        if isinstance(starts, (list, tuple)):
-            starts = [
-                item.numpy().item(0)
-                if isinstance(item, tmp_tensor_type)
-                else item
-                for item in starts
-            ]
-        elif isinstance(starts, tmp_tensor_type):
-            tensor_t = starts.numpy()
-            starts = [ele for ele in tensor_t]
-
-        if isinstance(ends, (list, tuple)):
-            ends = [
-                item.numpy().item(0)
-                if isinstance(item, tmp_tensor_type)
-                else item
-                for item in ends
-            ]
-            attrs += ('ends', ends)
-        elif isinstance(ends, tmp_tensor_type):
-            tensor_t = ends.numpy()
-            ends = [ele for ele in tensor_t]
-
-        return _C_ops.slice(input, axes, starts, ends, infer_flags, [])
-    else:
-        if _in_legacy_dygraph():
-            attrs = ()
-            starts_tensor = None
-            ends_tensor = None
-
-            if isinstance(axes, (list, tuple)):
-                axes = list(axes)
-                if len(axes) == 0:
-                    raise ValueError(
-                        "Input axes should not be an empty list/tuple."
-                    )
-                for i in range(len(axes)):
-                    if axes[i] < 0:
-                        axes[i] = max(0, axes[i] + len(input.shape))
-                    else:
-                        axes[i] = min(len(input.shape) - 1, axes[i])
-
-            else:
-                raise ValueError(
-                    "Input axes must be a python list or tuple, but reveived {}".format(
-                        type(axes)
-                    )
-                )
-
-            infer_flags = list(1 for i in range(len(axes)))
-
-            tmp_tensor_type = Variable
-
-            if isinstance(starts, (list, tuple)):
-                starts = [
-                    item.numpy().item(0)
-                    if isinstance(item, tmp_tensor_type)
-                    else item
-                    for item in starts
-                ]
-                attrs += ('starts', starts)
-            elif isinstance(starts, tmp_tensor_type):
-                starts_tensor = starts
-                starts.stop_gradient = True
-                infer_flags = list(-1 for i in range(len(axes)))
-
-            if isinstance(ends, (list, tuple)):
-                ends = [
-                    item.numpy().item(0)
-                    if isinstance(item, tmp_tensor_type)
-                    else item
-                    for item in ends
-                ]
-                attrs += ('ends', ends)
-            elif isinstance(ends, tmp_tensor_type):
-                ends_tensor = ends
-                ends_tensor.stop_gradient = True
-                infer_flags = list(-1 for i in range(len(axes)))
-
-            return _legacy_C_ops.slice(
-                input,
-                starts_tensor,
-                ends_tensor,
-                None,
-                None,
-                'axes',
-                axes,
-                'infer_flags',
-                infer_flags,
-                *attrs,
-            )
-
-    if not isinstance(starts, (list, tuple, Variable)):
-        raise ValueError(
-            "Input starts must be an Variable, python list or tuple."
-        )
-    if not isinstance(ends, (list, tuple, Variable)):
-        raise ValueError(
-            "Input ends must be an Variable, python list or tuple."
-        )
-
-    helper = LayerHelper('slice', **locals())
-
-    inputs = {'Input': input}
-    attrs = {'axes': axes}
-    infer_flags = list(1 for i in range(len(axes)))
-
-    # starts
-    if isinstance(starts, Variable):
-        starts.stop_gradient = True
-        inputs['StartsTensor'] = starts
-        infer_flags = list(-1 for i in range(len(axes)))
-    elif isinstance(starts, (list, tuple)):
-        attrs['starts'] = []
-        if utils._contain_var(starts):
-            inputs['StartsTensorList'] = utils._convert_to_tensor_list(starts)
-            for i, dim in enumerate(starts):
-                if isinstance(dim, Variable):
-                    attrs['starts'].append(-1)
-                    infer_flags[i] = -1
-                else:
-                    attrs['starts'].append(dim)
-        else:
-            attrs['starts'] = starts
-
-    # ends
-    if isinstance(ends, Variable):
-        ends.stop_gradient = True
-        inputs['EndsTensor'] = ends
-        infer_flags = list(-1 for i in range(len(axes)))
-    elif isinstance(ends, (list, tuple)):
-        attrs['ends'] = []
-        if utils._contain_var(ends):
-            inputs['EndsTensorList'] = utils._convert_to_tensor_list(ends)
-            for i, dim in enumerate(ends):
-                if isinstance(dim, Variable):
-                    attrs['ends'].append(-1)
-                    infer_flags[i] = -1
-                else:
-                    attrs['ends'].append(dim)
-        else:
-            attrs['ends'] = ends
-
-    # infer_flags
-    attrs['infer_flags'] = infer_flags
-    out = helper.create_variable_for_type_inference(
-        dtype=helper.input_dtype('input')
-    )
-    helper.append_op(
-        type='slice', inputs=inputs, attrs=attrs, outputs={'Out': out}
-    )
-
-    return out
-
-
 def shape(input):
     """
     :alias_main: paddle.shape

python/paddle/fluid/layers/rnn.py

@@ -2089,7 +2089,7 @@ class TrainingHelper(DecodeHelper):
         def _slice(x):  # TODO: use Variable.__getitem__
             axes = [0 if self.time_major else 1]
             return paddle.squeeze(
-                nn.slice(
+                paddle.slice(
                     x, axes=axes, starts=[next_time], ends=[next_time + 1]
                 ),
                 axis=axes,

python/paddle/fluid/tests/unittests/dygraph_to_static/bert_dygraph_model.py

@@ -291,7 +291,7 @@ class BertModelLayer(Layer):
         #
         # if not self.return_pooled_out:
         #    return enc_output
-        next_sent_feat = fluid.layers.slice(
+        next_sent_feat = paddle.slice(
             input=enc_output, axes=[1], starts=[0], ends=[1]
         )
         next_sent_feat = self.pooled_fc(next_sent_feat)

python/paddle/fluid/tests/unittests/dygraph_to_static/test_bmn.py

@@ -412,10 +412,10 @@ def bmn_loss_func(
         return loss
 
     pred_bm_reg = paddle.squeeze(
-        fluid.layers.slice(pred_bm, axes=[1], starts=[0], ends=[1]), axis=[1]
+        paddle.slice(pred_bm, axes=[1], starts=[0], ends=[1]), axis=[1]
     )
     pred_bm_cls = paddle.squeeze(
-        fluid.layers.slice(pred_bm, axes=[1], starts=[1], ends=[2]), axis=[1]
+        paddle.slice(pred_bm, axes=[1], starts=[1], ends=[2]), axis=[1]
     )
 
     bm_mask = _get_mask(cfg)

python/paddle/fluid/tests/unittests/dygraph_to_static/test_lac.py

@@ -85,9 +85,7 @@ class DynamicGRU(fluid.dygraph.Layer):
                 j = i
 
             # input_ = inputs[:, j:j+1, :]  # original code
-            input_ = fluid.layers.slice(
-                inputs, axes=[1], starts=[j], ends=[j + 1]
-            )
+            input_ = paddle.slice(inputs, axes=[1], starts=[j], ends=[j + 1])
             input_ = paddle.reshape(input_, [-1, input_.shape[2]])
             hidden, reset, gate = self.gru_unit(input_, hidden)
             hidden_ = paddle.reshape(hidden, [-1, 1, hidden.shape[1]])

python/paddle/fluid/tests/unittests/dygraph_to_static/yolov3.py

@@ -206,9 +206,7 @@ class Upsample(fluid.dygraph.Layer):
     def forward(self, inputs):
         # get dynamic upsample output shape
         shape_nchw = fluid.layers.shape(inputs)
-        shape_hw = fluid.layers.slice(
-            shape_nchw, axes=[0], starts=[2], ends=[4]
-        )
+        shape_hw = paddle.slice(shape_nchw, axes=[0], starts=[2], ends=[4])
         shape_hw.stop_gradient = True
         in_shape = fluid.layers.cast(shape_hw, dtype='int32')
         out_shape = in_shape * self.scale

python/paddle/fluid/tests/unittests/ipu/test_slice_op_ipu.py

@@ -51,7 +51,7 @@ class TestBase(IPUOpTest):
         x = paddle.static.data(
             name=self.feed_list[0], shape=self.feed_shape[0], dtype='float32'
         )
-        out = paddle.fluid.layers.slice(x, **self.attrs)
+        out = paddle.slice(x, **self.attrs)
         self.fetch_list = [out.name]
 
     def run_model(self, exec_mode):
@@ -105,9 +105,7 @@ class TestCase2(TestBase):
         ends = paddle.static.data(
             name=self.feed_list[2], shape=self.feed_shape[2], dtype='int32'
         )
-        out = paddle.fluid.layers.slice(
-            x, starts=starts, ends=ends, **self.attrs
-        )
+        out = paddle.slice(x, starts=starts, ends=ends, **self.attrs)
         self.fetch_list = [out.name]
 
 

python/paddle/fluid/tests/unittests/ir/inference/test_trt_slice_dynamic_plugin.py

@@ -17,6 +17,7 @@ import unittest
 import numpy as np
 from inference_pass_test import InferencePassTest
 
+import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.core import AnalysisConfig
@@ -49,9 +50,7 @@ class SlicePluginTRTDynamicTest(InferencePassTest):
             axes = self.params_axes
             starts = self.params_starts
             ends = self.params_ends
-            slice_out = fluid.layers.slice(
-                data, axes=axes, starts=starts, ends=ends
-            )
+            slice_out = paddle.slice(data, axes=axes, starts=starts, ends=ends)
 
         self.feeds = {
             "data": np.random.random((3, 3, 3, 3)).astype("float32"),

python/paddle/fluid/tests/unittests/ir/inference/test_trt_slice_plugin.py

@@ -17,6 +17,7 @@ import unittest
 import numpy as np
 from inference_pass_test import InferencePassTest
 
+import paddle
 import paddle.fluid as fluid
 import paddle.fluid.core as core
 from paddle.fluid.core import AnalysisConfig
@@ -43,9 +44,7 @@ class SlicePluginTRTTest(InferencePassTest):
             axes = self.params_axes
             starts = self.params_starts
             ends = self.params_ends
-            slice_out = fluid.layers.slice(
-                data, axes=axes, starts=starts, ends=ends
-            )
+            slice_out = paddle.slice(data, axes=axes, starts=starts, ends=ends)
             out = fluid.layers.batch_norm(slice_out, is_test=True)
 
         self.feeds = {
@@ -114,9 +113,7 @@ class SlicePluginTRTTestInt32(SlicePluginTRTTest):
             axes = self.params_axes
             starts = self.params_starts
             ends = self.params_ends
-            slice_out = fluid.layers.slice(
-                data, axes=axes, starts=starts, ends=ends
-            )
+            slice_out = paddle.slice(data, axes=axes, starts=starts, ends=ends)
             cast_out = fluid.layers.cast(slice_out, 'float32')
             out = fluid.layers.batch_norm(cast_out, is_test=True)
 
@@ -141,9 +138,7 @@ class StaticSlicePluginTRTTestInt32(SlicePluginTRTTest):
             axes = self.params_axes
             starts = self.params_starts
             ends = self.params_ends
-            slice_out = fluid.layers.slice(
-                data, axes=axes, starts=starts, ends=ends
-            )
+            slice_out = paddle.slice(data, axes=axes, starts=starts, ends=ends)
             cast_out = fluid.layers.cast(slice_out, 'float32')
             out = fluid.layers.batch_norm(cast_out, is_test=True)
 

python/paddle/fluid/tests/unittests/test_eager_deletion_padding_rnn.py

@@ -143,10 +143,10 @@ def lm_model(
             )
             bias_arr.append(bias_1)
 
-            pre_hidden = layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(pre_hidden, shape=[-1, hidden_size])
@@ -169,22 +169,22 @@ def lm_model(
                 gate_input = layers.matmul(x=nn, y=weight_1)
 
                 gate_input = layers.elementwise_add(gate_input, bias)
-                i = layers.slice(
+                i = paddle.slice(
                     gate_input, axes=[1], starts=[0], ends=[hidden_size]
                 )
-                j = layers.slice(
+                j = paddle.slice(
                     gate_input,
                     axes=[1],
                     starts=[hidden_size],
                     ends=[hidden_size * 2],
                 )
-                f = layers.slice(
+                f = paddle.slice(
                     gate_input,
                     axes=[1],
                     starts=[hidden_size * 2],
                     ends=[hidden_size * 3],
                 )
-                o = layers.slice(
+                o = paddle.slice(
                     gate_input,
                     axes=[1],
                     starts=[hidden_size * 3],
@@ -222,11 +222,11 @@ def lm_model(
             c = rnnout[i * 2 + 1]
             m.stop_gradient = True
             c.stop_gradient = True
-            last_h = layers.slice(
+            last_h = paddle.slice(
                 m, axes=[0], starts=[num_steps - 1], ends=[num_steps]
             )
             last_hidden_array.append(last_h)
-            last_c = layers.slice(
+            last_c = paddle.slice(
                 c, axes=[0], starts=[num_steps - 1], ends=[num_steps]
             )
             last_cell_array.append(last_c)
@@ -264,10 +264,10 @@ def lm_model(
             )
             bias_arr.append(bias_1)
 
-            pre_hidden = layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(pre_hidden, shape=[-1, hidden_size])

python/paddle/fluid/tests/unittests/test_imperative_basic.py

@@ -137,9 +137,7 @@ class SimpleRNN(fluid.Layer):
         )
         pre_hidden = init_hidden
         for i in range(self.seq_len):
-            input = fluid.layers.slice(
-                inputs, axes=[1], starts=[i], ends=[i + 1]
-            )
+            input = paddle.slice(inputs, axes=[1], starts=[i], ends=[i + 1])
             input = paddle.reshape(input, shape=[1, 3])
             out_softmax, pre_hidden = self._cell(input, pre_hidden)
             outs.append(out_softmax)

python/paddle/fluid/tests/unittests/test_imperative_ocr_attention_model.py

@@ -192,9 +192,7 @@ class DynamicGRU(fluid.dygraph.Layer):
         for i in range(inputs.shape[1]):
             if self.is_reverse:
                 i = inputs.shape[1] - 1 - i
-            input_ = fluid.layers.slice(
-                inputs, axes=[1], starts=[i], ends=[i + 1]
-            )
+            input_ = paddle.slice(inputs, axes=[1], starts=[i], ends=[i + 1])
             input_ = paddle.reshape(input_, [-1, input_.shape[2]])
             hidden, reset, gate = self.gru_unit(input_, hidden)
             hidden_ = paddle.reshape(hidden, [-1, 1, hidden.shape[1]])
@@ -356,7 +354,7 @@ class GRUDecoderWithAttention(fluid.dygraph.Layer):
         res = []
         hidden_mem = decoder_boot
         for i in range(target_embedding.shape[1]):
-            current_word = fluid.layers.slice(
+            current_word = paddle.slice(
                 target_embedding, axes=[1], starts=[i], ends=[i + 1]
             )
             current_word = paddle.reshape(
@@ -399,7 +397,7 @@ class OCRAttention(fluid.dygraph.Layer):
 
     def forward(self, inputs, label_in):
         gru_backward, encoded_vector, encoded_proj = self.encoder_net(inputs)
-        backward_first = fluid.layers.slice(
+        backward_first = paddle.slice(
             gru_backward, axes=[1], starts=[0], ends=[1]
         )
         backward_first = paddle.reshape(

python/paddle/fluid/tests/unittests/test_imperative_ptb_rnn.py

@@ -79,10 +79,10 @@ class SimpleLSTMRNN(fluid.Layer):
         self.hidden_array = []
 
         for i in range(self._num_layers):
-            pre_hidden = fluid.layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = fluid.layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(
@@ -94,7 +94,7 @@ class SimpleLSTMRNN(fluid.Layer):
 
         res = []
         for index in range(self._num_steps):
-            self._input = fluid.layers.slice(
+            self._input = paddle.slice(
                 input_embedding, axes=[1], starts=[index], ends=[index + 1]
             )
             self._input = paddle.reshape(

python/paddle/fluid/tests/unittests/test_imperative_save_load.py

@@ -74,10 +74,10 @@ class SimpleLSTMRNN(fluid.Layer):
         self.hidden_array = []
 
         for i in range(self._num_layers):
-            pre_hidden = fluid.layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = fluid.layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(
@@ -89,7 +89,7 @@ class SimpleLSTMRNN(fluid.Layer):
 
         res = []
         for index in range(self._num_steps):
-            self._input = fluid.layers.slice(
+            self._input = paddle.slice(
                 input_embedding, axes=[1], starts=[index], ends=[index + 1]
             )
             self._input = paddle.reshape(

python/paddle/fluid/tests/unittests/test_imperative_save_load_v2.py

@@ -77,10 +77,10 @@ class SimpleLSTMRNN(fluid.Layer):
         self.hidden_array = []
 
         for i in range(self._num_layers):
-            pre_hidden = fluid.layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = fluid.layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(
@@ -92,7 +92,7 @@ class SimpleLSTMRNN(fluid.Layer):
 
         res = []
         for index in range(self._num_steps):
-            self._input = fluid.layers.slice(
+            self._input = paddle.slice(
                 input_embedding, axes=[1], starts=[index], ends=[index + 1]
             )
             self._input = paddle.reshape(

python/paddle/fluid/tests/unittests/test_layers.py

@@ -3584,7 +3584,7 @@ class TestBook(LayerTest):
                 name="input", shape=[3, 4, 5, 6], dtype='float32'
             )
 
-            out = layers.slice(input, axes=axes, starts=starts, ends=ends)
+            out = paddle.slice(input, axes=axes, starts=starts, ends=ends)
             return out
 
     def make_scale_variable(self):

python/paddle/fluid/tests/unittests/test_nn_grad.py

@@ -30,7 +30,7 @@ class TestSliceOpDoubleGradCheck(unittest.TestCase):
     def func(self, place):
         self.config()
 
-        out = fluid.layers.slice(
+        out = paddle.slice(
             self.inputs, axes=self.axes, starts=self.starts, ends=self.ends
         )
         gradient_checker.double_grad_check(

python/paddle/fluid/tests/unittests/test_static_save_load.py

@@ -85,10 +85,10 @@ class SimpleLSTMRNN(fluid.Layer):
         self.hidden_array = []
 
         for i in range(self._num_layers):
-            pre_hidden = fluid.layers.slice(
+            pre_hidden = paddle.slice(
                 init_hidden, axes=[0], starts=[i], ends=[i + 1]
             )
-            pre_cell = fluid.layers.slice(
+            pre_cell = paddle.slice(
                 init_cell, axes=[0], starts=[i], ends=[i + 1]
             )
             pre_hidden = paddle.reshape(
@@ -100,7 +100,7 @@ class SimpleLSTMRNN(fluid.Layer):
 
         res = []
         for index in range(self._num_steps):
-            self._input = fluid.layers.slice(
+            self._input = paddle.slice(
                 input_embedding, axes=[1], starts=[index], ends=[index + 1]
             )
             self._input = paddle.reshape(